1. CS 416 Artificial Intelligence Lecture 5 Finish Uninformed Searches Begin Informed Searches Lecture 5 Finish Uninformed Searches Begin Informed Searches

2. Uniform-cost search (review) Always expand the lowest-path-cost node Don’t evaluate the node until it is “expanded”… not when it is the result of expansion Always expand the lowest-path-cost node Don’t evaluate the node until it is “expanded”… not when it is the result of expansion

3. Uniform-cost search (review) Fringe = [S 0 ] Expand(S)  {A 1, B 5, C 15 } Fringe = [S 0 ] Expand(S)  {A 1, B 5, C 15 } B

4. Uniform-cost search (review) Fringe = [A 1, B 5, C 15 ] Expand(A) = {G 11 } Fringe = [A 1, B 5, C 15 ] Expand(A) = {G 11 }

5. Uniform-cost search (review) Fringe = [B 5, G 11, C 15 ] Expand(B)  {G 10 } Fringe = [B 5, G 11, C 15 ] Expand(B)  {G 10 }

6. Uniform-cost search (review) Fringe = [G 10, C 15 ] Expand(G)  Goal Fringe = [G 10, C 15 ] Expand(G)  Goal

7. Bidirectional search Search from goal to start Search from start to goal Search from goal to start Search from start to goal Two b d/2 searches instead of one b d

8. Bidirectional search Implementation Each search checks nodes before expansion to see if they are on the fringe of the other searchEach search checks nodes before expansion to see if they are on the fringe of the other searchImplementation

9. Bidirectional search Example: Bidirectional BFS search with d=6 & b=10Example: Bidirectional BFS search with d=6 & b=10 –Worst case: both search trees must expand all but one element of the third level of the tree  2* (1 + 10 + 100 + 1000 + 10000 - 10) node expansions  Versus 1 * (1 + 10 + 100 + … + 10000000) expansions Example: Bidirectional BFS search with d=6 & b=10Example: Bidirectional BFS search with d=6 & b=10 –Worst case: both search trees must expand all but one element of the third level of the tree  2* (1 + 10 + 100 + 1000 + 10000 - 10) node expansions  Versus 1 * (1 + 10 + 100 + … + 10000000) expansions

10. Bidirectional search Implementation Checking fringe of other treeChecking fringe of other tree –At least one search tree must be kept in memory –Checking can be done in constant time (hash table) Searching back from goalSearching back from goal –Must be able to compute predecessors to node n: Pred (n) –Easy with 15-puzzle, but how about chess? Implementation Checking fringe of other treeChecking fringe of other tree –At least one search tree must be kept in memory –Checking can be done in constant time (hash table) Searching back from goalSearching back from goal –Must be able to compute predecessors to node n: Pred (n) –Easy with 15-puzzle, but how about chess?

11. Avoid repeated states Search algorithms that forget their history are doomed to repeat it. Russell and Norvig So remember where you’ve been… on a listSo remember where you’ve been… on a list If you come upon a node you’ve visited before, don’t expand itIf you come upon a node you’ve visited before, don’t expand it Let’s call this GRAPH-SEARCHLet’s call this GRAPH-SEARCH Search algorithms that forget their history are doomed to repeat it. Russell and Norvig So remember where you’ve been… on a listSo remember where you’ve been… on a list If you come upon a node you’ve visited before, don’t expand itIf you come upon a node you’ve visited before, don’t expand it Let’s call this GRAPH-SEARCHLet’s call this GRAPH-SEARCH

12. GRAPH-SEARCH Faster and smaller space requirements if many repeating states Time-space requirements are a function of state space not depth/branching of tree to goalTime-space requirements are a function of state space not depth/branching of tree to goal At this point in class, repeated states will be thrown away – even if the new path to the state is better than the one explored previouslyAt this point in class, repeated states will be thrown away – even if the new path to the state is better than the one explored previously Faster and smaller space requirements if many repeating states Time-space requirements are a function of state space not depth/branching of tree to goalTime-space requirements are a function of state space not depth/branching of tree to goal At this point in class, repeated states will be thrown away – even if the new path to the state is better than the one explored previouslyAt this point in class, repeated states will be thrown away – even if the new path to the state is better than the one explored previously

13. Interesting problems Exercise 3.9: 3 cannibals and 3 missionaries and a boat that can hold one or two people are on one side of the river. Get everyone across the river (early AI problem, 1968)3 cannibals and 3 missionaries and a boat that can hold one or two people are on one side of the river. Get everyone across the river (early AI problem, 1968) 8-puzzle and 15-puzzle, invented by Sam Loyd in good ol’ USA in 1870s. Think about search space.8-puzzle and 15-puzzle, invented by Sam Loyd in good ol’ USA in 1870s. Think about search space. Rubik’s cubeRubik’s cube Traveling Salesman Problem (TSP)Traveling Salesman Problem (TSP) Exercise 3.9: 3 cannibals and 3 missionaries and a boat that can hold one or two people are on one side of the river. Get everyone across the river (early AI problem, 1968)3 cannibals and 3 missionaries and a boat that can hold one or two people are on one side of the river. Get everyone across the river (early AI problem, 1968) 8-puzzle and 15-puzzle, invented by Sam Loyd in good ol’ USA in 1870s. Think about search space.8-puzzle and 15-puzzle, invented by Sam Loyd in good ol’ USA in 1870s. Think about search space. Rubik’s cubeRubik’s cube Traveling Salesman Problem (TSP)Traveling Salesman Problem (TSP)

14. Chapter 4 – Informed Search INFORMED? Uses problem-specific knowledge beyond the definition of the problem itselfUses problem-specific knowledge beyond the definition of the problem itself –selecting best lane in traffic –playing a sport  what’s the heuristic (or evaluation function)? INFORMED? Uses problem-specific knowledge beyond the definition of the problem itselfUses problem-specific knowledge beyond the definition of the problem itself –selecting best lane in traffic –playing a sport  what’s the heuristic (or evaluation function)? www.curling.ca

15. Best-first search BFS/DFS/UCS differ in how they select a node to pull off the fringe We want to pull off the fringe the node that’s on the optimal pathWe want to pull off the fringe the node that’s on the optimal path –But if we know the “best” node to explore, we don’t have to search!!! BFS/DFS/UCS differ in how they select a node to pull off the fringe We want to pull off the fringe the node that’s on the optimal pathWe want to pull off the fringe the node that’s on the optimal path –But if we know the “best” node to explore, we don’t have to search!!! We’re not certain we know the best node

16. Best-first search Use an evaluation function to select node to expand f(n) = evaluation function = expected “cost” for a path from root to goal that goes through node nf(n) = evaluation function = expected “cost” for a path from root to goal that goes through node n Select the node n on fringe that minimizes f(n)Select the node n on fringe that minimizes f(n) How do we build f(n)?How do we build f(n)? Use an evaluation function to select node to expand f(n) = evaluation function = expected “cost” for a path from root to goal that goes through node nf(n) = evaluation function = expected “cost” for a path from root to goal that goes through node n Select the node n on fringe that minimizes f(n)Select the node n on fringe that minimizes f(n) How do we build f(n)?How do we build f(n)?

17. Evaluation function: f(n) Combine two costs f(n) = g(n) + h(n)f(n) = g(n) + h(n) –g(n) = cost to get to n from start  We know this! –h(n) = cost to get to from n to goal Combine two costs f(n) = g(n) + h(n)f(n) = g(n) + h(n) –g(n) = cost to get to n from start  We know this! –h(n) = cost to get to from n to goal

18. Heuristics A function, h(n), that estimates cost of cheapest path from node n to the goal h(n) = 0 if n == goal nodeh(n) = 0 if n == goal node A function, h(n), that estimates cost of cheapest path from node n to the goal h(n) = 0 if n == goal nodeh(n) = 0 if n == goal node

19. Greedy best-first search Trust your heuristic and ignore path costs evaluate node that minimizes h(n)evaluate node that minimizes h(n) f(n) = h(n)f(n) = h(n) Example: getting from A to B Explore nodes with shortest straight distance to BExplore nodes with shortest straight distance to B Shortcomings of heuristic?Shortcomings of heuristic? –Greedy can be bad Trust your heuristic and ignore path costs evaluate node that minimizes h(n)evaluate node that minimizes h(n) f(n) = h(n)f(n) = h(n) Example: getting from A to B Explore nodes with shortest straight distance to BExplore nodes with shortest straight distance to B Shortcomings of heuristic?Shortcomings of heuristic? –Greedy can be bad

20. A* (A-star) Search Don’t simply minimize the cost to goal… minimize the cost from start to goal… f(n) = g(n) + h(n)f(n) = g(n) + h(n) –g(n) = cost to get to n from start –h(n) = cost to get from n to goal Select node from fringe that minimizes f(n) Don’t simply minimize the cost to goal… minimize the cost from start to goal… f(n) = g(n) + h(n)f(n) = g(n) + h(n) –g(n) = cost to get to n from start –h(n) = cost to get from n to goal Select node from fringe that minimizes f(n)

21. A* is Optimal? A* can be optimal if h(n) satisfies conditions h(n) never overestimates cost to reach the goalh(n) never overestimates cost to reach the goal –it is eternally optimistic –called an admissible heuristic –f(n) never overestimates cost of a solution through n Proof of optimality?Proof of optimality? A* can be optimal if h(n) satisfies conditions h(n) never overestimates cost to reach the goalh(n) never overestimates cost to reach the goal –it is eternally optimistic –called an admissible heuristic –f(n) never overestimates cost of a solution through n Proof of optimality?Proof of optimality?

22. A* is Optimal We must prove that A* will not return a suboptimal goal or a suboptimal path to a goal Let G be a suboptimal goal nodeLet G be a suboptimal goal node –f(G) = g(G) + h(G) –h(G) = 0 because G is a goal node and we cannot overestimate actual cost to reach G –f(G) = g(G) > C* (because G is suboptimal) We must prove that A* will not return a suboptimal goal or a suboptimal path to a goal Let G be a suboptimal goal nodeLet G be a suboptimal goal node –f(G) = g(G) + h(G) –h(G) = 0 because G is a goal node and we cannot overestimate actual cost to reach G –f(G) = g(G) > C* (because G is suboptimal)

23. A* is Optimal (cont.) Let n be a node on the optimal pathLet n be a node on the optimal path –because h(n) does not overestimate –f(n) = g(n) + h(n) <= C* Therefore f(n) <= C* < f(G)Therefore f(n) <= C* < f(G) –node n will be selected before node G Let n be a node on the optimal pathLet n be a node on the optimal path –because h(n) does not overestimate –f(n) = g(n) + h(n) <= C* Therefore f(n) <= C* < f(G)Therefore f(n) <= C* < f(G) –node n will be selected before node G A* is optimal if h() is admissible

24. Repeated States and GRAPH-SEARCH GRAPH-SEARCH always ignores all but the first occurrence of a state during search Lower cost path may be tossedLower cost path may be tossed –So, don’t throw away subsequent occurrences –Or, ensure that the optimal path to any repeated state is always the first one followed Additional constraint on heurisitic, consistencyAdditional constraint on heurisitic, consistency GRAPH-SEARCH always ignores all but the first occurrence of a state during search Lower cost path may be tossedLower cost path may be tossed –So, don’t throw away subsequent occurrences –Or, ensure that the optimal path to any repeated state is always the first one followed Additional constraint on heurisitic, consistencyAdditional constraint on heurisitic, consistency

25. Consistent heuristic: h(n) Heuristic function must be monotonic for every node, n, and successor, n’, obtained with action afor every node, n, and successor, n’, obtained with action a –estimated cost of reaching goal from n is no greater than cost of getting to n’ plus estimated cost of reaching goal from n’ –h(n) <= c(n, a, n’) + h(n’) This implies f(n) along any path are nondecreasingThis implies f(n) along any path are nondecreasing Heuristic function must be monotonic for every node, n, and successor, n’, obtained with action afor every node, n, and successor, n’, obtained with action a –estimated cost of reaching goal from n is no greater than cost of getting to n’ plus estimated cost of reaching goal from n’ –h(n) <= c(n, a, n’) + h(n’) This implies f(n) along any path are nondecreasingThis implies f(n) along any path are nondecreasing

26. Examples of consistent h(n) h(n) <= c(n, a, n succ ) + h(n succ ) recall h(n) is admissiblerecall h(n) is admissible –The quickest you can get there from here is 10 minutes  It may take more than 10 minutes, but not fewer After taking an action and learning the costAfter taking an action and learning the cost –It took you two minutes to get here and you still have nine minutes to go –We cannot learn… it took you two minutes to get here and you have seven minutes to go h(n) <= c(n, a, n succ ) + h(n succ ) recall h(n) is admissiblerecall h(n) is admissible –The quickest you can get there from here is 10 minutes  It may take more than 10 minutes, but not fewer After taking an action and learning the costAfter taking an action and learning the cost –It took you two minutes to get here and you still have nine minutes to go –We cannot learn… it took you two minutes to get here and you have seven minutes to go 2 10 9 0

27. Example of inconsistent h(n) As a thought exercise for after class Consider what happens when a heuristic is inconsistentConsider what happens when a heuristic is inconsistent Consider how one could have a consistent but non- admissible heuristicConsider how one could have a consistent but non- admissible heuristic As a thought exercise for after class Consider what happens when a heuristic is inconsistentConsider what happens when a heuristic is inconsistent Consider how one could have a consistent but non- admissible heuristicConsider how one could have a consistent but non- admissible heuristic

28. Proof of monotonicity of f(n) If h(n) is consistent (monotonic) then f(n) along any path is nondecreasing then f(n) along any path is nondecreasing let n’ be a successor of nlet n’ be a successor of n –g(n’) = g(n) + c (n, a, n’) for some a –f(n’) = g(n’) + h(n’) = g(n) + c(n, a, n’) + h(n’) >= g(n) + h(n) = f(n) If h(n) is consistent (monotonic) then f(n) along any path is nondecreasing then f(n) along any path is nondecreasing let n’ be a successor of nlet n’ be a successor of n –g(n’) = g(n) + c (n, a, n’) for some a –f(n’) = g(n’) + h(n’) = g(n) + c(n, a, n’) + h(n’) >= g(n) + h(n) = f(n) monotonicity implies h(n) <= c(n, a, n’) + h(n’)

29. Contours Because f(n) is nondecreasing we can draw contours If we know C*If we know C* We only need to explore contours less than C*We only need to explore contours less than C* Because f(n) is nondecreasing we can draw contours If we know C*If we know C* We only need to explore contours less than C*We only need to explore contours less than C*

30. Properties of A* A* expands all nodes with f(n) < C*A* expands all nodes with f(n) < C* A* expands some (at least one) of the nodes on the C* contour before finding the goalA* expands some (at least one) of the nodes on the C* contour before finding the goal A* expands no nodes with f(n) > C*A* expands no nodes with f(n) > C* –these unexpanded nodes can be pruned A* expands all nodes with f(n) < C*A* expands all nodes with f(n) < C* A* expands some (at least one) of the nodes on the C* contour before finding the goalA* expands some (at least one) of the nodes on the C* contour before finding the goal A* expands no nodes with f(n) > C*A* expands no nodes with f(n) > C* –these unexpanded nodes can be pruned

31. A* is Optimally Efficient Compared to other algorithms that search from root Compared to other algorithms using same heuristic No other optimal algorithm is guaranteed to expand fewer nodes than A* (except perhaps eliminating consideration of ties at f(n) = C*) Compared to other algorithms that search from root Compared to other algorithms using same heuristic No other optimal algorithm is guaranteed to expand fewer nodes than A* (except perhaps eliminating consideration of ties at f(n) = C*)

32. Pros and Cons of A* A* is optimal and optimally efficient A* is still slow and bulky (space kills first) Number of nodes grows exponentially with the length to goalNumber of nodes grows exponentially with the length to goal –This is actually a function of heuristic, but they all have errors A* must search all nodes within this goal contourA* must search all nodes within this goal contour Finding suboptimal goals is sometimes only feasible solutionFinding suboptimal goals is sometimes only feasible solution Sometimes, better heuristics are non-admissibleSometimes, better heuristics are non-admissible A* is optimal and optimally efficient A* is still slow and bulky (space kills first) Number of nodes grows exponentially with the length to goalNumber of nodes grows exponentially with the length to goal –This is actually a function of heuristic, but they all have errors A* must search all nodes within this goal contourA* must search all nodes within this goal contour Finding suboptimal goals is sometimes only feasible solutionFinding suboptimal goals is sometimes only feasible solution Sometimes, better heuristics are non-admissibleSometimes, better heuristics are non-admissible

33. Memory-bounded Heuristic Search Try to reduce memory needs Take advantage of heuristic to improve performance Iterative-deepening A* (IDA*)Iterative-deepening A* (IDA*) Recursive best-first search (RBFS)Recursive best-first search (RBFS) SMA*SMA* Try to reduce memory needs Take advantage of heuristic to improve performance Iterative-deepening A* (IDA*)Iterative-deepening A* (IDA*) Recursive best-first search (RBFS)Recursive best-first search (RBFS) SMA*SMA*

34. Iterative Deepening A* Iterative Deepening Remember from uninformed search, this was a depth-first search where the max depth was iteratively increasedRemember from uninformed search, this was a depth-first search where the max depth was iteratively increased As an informed search, we again perform depth-first search, but only nodes with f-cost less than or equal to smallest f- cost of nodes expanded at last iterationAs an informed search, we again perform depth-first search, but only nodes with f-cost less than or equal to smallest f- cost of nodes expanded at last iteration –What happens when f-cost is real-valued? Iterative Deepening Remember from uninformed search, this was a depth-first search where the max depth was iteratively increasedRemember from uninformed search, this was a depth-first search where the max depth was iteratively increased As an informed search, we again perform depth-first search, but only nodes with f-cost less than or equal to smallest f- cost of nodes expanded at last iterationAs an informed search, we again perform depth-first search, but only nodes with f-cost less than or equal to smallest f- cost of nodes expanded at last iteration –What happens when f-cost is real-valued?

35. Recursive best-first search Depth-first combined with best alternative Keep track of options along fringeKeep track of options along fringe As soon as current depth-first exploration becomes more expensive of best fringe optionAs soon as current depth-first exploration becomes more expensive of best fringe option –back up to fringe, but update node costs along the way Depth-first combined with best alternative Keep track of options along fringeKeep track of options along fringe As soon as current depth-first exploration becomes more expensive of best fringe optionAs soon as current depth-first exploration becomes more expensive of best fringe option –back up to fringe, but update node costs along the way

36. Recursive best-first search

39. Quality of Iterative Deepening A* and Recursive best-first search RBFS O(bd) space complexity [if h(n) is admissible]O(bd) space complexity [if h(n) is admissible] Time complexity is hard to describeTime complexity is hard to describe –efficiency is heavily dependent on quality of h(n) –same states may be explored many times IDA* and RBFS use too little memoryIDA* and RBFS use too little memory –even if you wanted to use more than O(bd) memory, these two could not provide any advantage RBFS O(bd) space complexity [if h(n) is admissible]O(bd) space complexity [if h(n) is admissible] Time complexity is hard to describeTime complexity is hard to describe –efficiency is heavily dependent on quality of h(n) –same states may be explored many times IDA* and RBFS use too little memoryIDA* and RBFS use too little memory –even if you wanted to use more than O(bd) memory, these two could not provide any advantage

40. Simple Memory-bounded A* Use all available memory Follow A* algorithm and fill memory with new expanded nodesFollow A* algorithm and fill memory with new expanded nodes If new node does not fitIf new node does not fit –free() stored node with worst f-value –propagate f-value of freed node to parent SMA* will regenerate a subtree only when it is neededSMA* will regenerate a subtree only when it is needed –the path through deleted subtree is unknown, but cost is known Use all available memory Follow A* algorithm and fill memory with new expanded nodesFollow A* algorithm and fill memory with new expanded nodes If new node does not fitIf new node does not fit –free() stored node with worst f-value –propagate f-value of freed node to parent SMA* will regenerate a subtree only when it is neededSMA* will regenerate a subtree only when it is needed –the path through deleted subtree is unknown, but cost is known

41. Thrashing Typically discussed in OS w.r.t. memory The cost of repeatedly freeing and regenerating parts of the search tree dominate the cost of actual searchThe cost of repeatedly freeing and regenerating parts of the search tree dominate the cost of actual search time complexity will scale significantly if thrashingtime complexity will scale significantly if thrashing –So we saved space with SMA*, but if the problem is large, it will be intractable from the point of view of computation time Typically discussed in OS w.r.t. memory The cost of repeatedly freeing and regenerating parts of the search tree dominate the cost of actual searchThe cost of repeatedly freeing and regenerating parts of the search tree dominate the cost of actual search time complexity will scale significantly if thrashingtime complexity will scale significantly if thrashing –So we saved space with SMA*, but if the problem is large, it will be intractable from the point of view of computation time

42. Meta-foo What does meta mean in AI? Frequently it means step back a level from fooFrequently it means step back a level from foo Metareasoning = reasoning about reasoningMetareasoning = reasoning about reasoning These informed search algorithms have pros and cons regarding how they choose to explore new levelsThese informed search algorithms have pros and cons regarding how they choose to explore new levels –a metalevel learning algorithm may combine learn how to combine techniques and parameterize search What does meta mean in AI? Frequently it means step back a level from fooFrequently it means step back a level from foo Metareasoning = reasoning about reasoningMetareasoning = reasoning about reasoning These informed search algorithms have pros and cons regarding how they choose to explore new levelsThese informed search algorithms have pros and cons regarding how they choose to explore new levels –a metalevel learning algorithm may combine learn how to combine techniques and parameterize search

43. Heuristic Functions 8-puzzle problem Avg Depth=22 Branching = approx 3 3 22 states 170,000 repeated

44. Heuristics The number of misplaced tiles (h 1 ) Admissible because at least n moves required to solve n misplaced tilesAdmissible because at least n moves required to solve n misplaced tiles The distance from each tile to its goal position (h 2 ) No diagonals, so use Manhattan DistanceNo diagonals, so use Manhattan Distance –As if walking around rectilinear city blocks also admissiblealso admissible The number of misplaced tiles (h 1 ) Admissible because at least n moves required to solve n misplaced tilesAdmissible because at least n moves required to solve n misplaced tiles The distance from each tile to its goal position (h 2 ) No diagonals, so use Manhattan DistanceNo diagonals, so use Manhattan Distance –As if walking around rectilinear city blocks also admissiblealso admissible

45. Compare these two heuristics Effective Branching Factor, b* If A* creates N nodes to find the goal at depth dIf A* creates N nodes to find the goal at depth d –b* = branching factor such that a uniform tree of depth d contains N+1 nodes (add 1 to N to account for root)  N+1 = 1 + b* + (b*) 2 + … + (b*) d b* close to 1 is idealb* close to 1 is ideal –because this means the heuristic guided the A* search linearly –If b* were 100, on average, the heuristic had to consider 100 children for each node –Compare heuristics based on their b* Effective Branching Factor, b* If A* creates N nodes to find the goal at depth dIf A* creates N nodes to find the goal at depth d –b* = branching factor such that a uniform tree of depth d contains N+1 nodes (add 1 to N to account for root)  N+1 = 1 + b* + (b*) 2 + … + (b*) d b* close to 1 is idealb* close to 1 is ideal –because this means the heuristic guided the A* search linearly –If b* were 100, on average, the heuristic had to consider 100 children for each node –Compare heuristics based on their b*

46. Compare these two heuristics

47. h 2 is always better than h 1 for any node, n, h 2 (n) >= h 1 (n)for any node, n, h 2 (n) >= h 1 (n) h 2 dominates h 1h 2 dominates h 1 Recall all nodes with f(n) < C* will be expanded?Recall all nodes with f(n) < C* will be expanded? –This means all nodes, h(n) + g(n) < C*, will be expanded  All nodes where h(n) < C* - g(n) will be expanded –All nodes h 2 expands will also be expanded by h 1 and because h 1 is smaller, others will be expanded as well h 2 is always better than h 1 for any node, n, h 2 (n) >= h 1 (n)for any node, n, h 2 (n) >= h 1 (n) h 2 dominates h 1h 2 dominates h 1 Recall all nodes with f(n) < C* will be expanded?Recall all nodes with f(n) < C* will be expanded? –This means all nodes, h(n) + g(n) < C*, will be expanded  All nodes where h(n) < C* - g(n) will be expanded –All nodes h 2 expands will also be expanded by h 1 and because h 1 is smaller, others will be expanded as well

48. Inventing admissible heuristic funcs How can you create h(n)? Simplify problem by reducing restrictions on actionsSimplify problem by reducing restrictions on actions –Allow 8-puzzle pieces to sit atop on another –Call this a relaxed problem –The cost of optimal solution to relaxed problem is admissible heuristic for original problem  It is at least as expensive for the original problem How can you create h(n)? Simplify problem by reducing restrictions on actionsSimplify problem by reducing restrictions on actions –Allow 8-puzzle pieces to sit atop on another –Call this a relaxed problem –The cost of optimal solution to relaxed problem is admissible heuristic for original problem  It is at least as expensive for the original problem

49. Examples of relaxed problems A tile can move from square A to square B if A is horizontally or vertically adjacent to B and B is blank A tile can move from A to B if A is adjacent to B (overlap)A tile can move from A to B if A is adjacent to B (overlap) A tile can move from A to B if B is blank (teleport)A tile can move from A to B if B is blank (teleport) A tile can move from A to B (teleport and overlap)A tile can move from A to B (teleport and overlap) Solutions to these relaxed problems can be computed without search and therefore heuristic is easy to compute A tile can move from square A to square B if A is horizontally or vertically adjacent to B and B is blank A tile can move from A to B if A is adjacent to B (overlap)A tile can move from A to B if A is adjacent to B (overlap) A tile can move from A to B if B is blank (teleport)A tile can move from A to B if B is blank (teleport) A tile can move from A to B (teleport and overlap)A tile can move from A to B (teleport and overlap) Solutions to these relaxed problems can be computed without search and therefore heuristic is easy to compute

50. Multiple Heuristics If multiple heuristics available: h(n) = max {h 1 (n), h 2 (n), …, h m (n)}h(n) = max {h 1 (n), h 2 (n), …, h m (n)} If multiple heuristics available: h(n) = max {h 1 (n), h 2 (n), …, h m (n)}h(n) = max {h 1 (n), h 2 (n), …, h m (n)}

51. Use solution to subproblem as heuristic What is optimal cost of solving some portion of original problem? subproblem solution is heuristic of original problemsubproblem solution is heuristic of original problem What is optimal cost of solving some portion of original problem? subproblem solution is heuristic of original problemsubproblem solution is heuristic of original problem

52. Pattern Databases Store optimal solutions to subproblems in database We use an exhaustive search to solve every permutation of the 1,2,3,4-piece subproblem of the 8-puzzleWe use an exhaustive search to solve every permutation of the 1,2,3,4-piece subproblem of the 8-puzzle During solution of 8-puzzle, look up optimal cost to solve the 1,2,3,4-piece subproblem and use as heuristicDuring solution of 8-puzzle, look up optimal cost to solve the 1,2,3,4-piece subproblem and use as heuristic Store optimal solutions to subproblems in database We use an exhaustive search to solve every permutation of the 1,2,3,4-piece subproblem of the 8-puzzleWe use an exhaustive search to solve every permutation of the 1,2,3,4-piece subproblem of the 8-puzzle During solution of 8-puzzle, look up optimal cost to solve the 1,2,3,4-piece subproblem and use as heuristicDuring solution of 8-puzzle, look up optimal cost to solve the 1,2,3,4-piece subproblem and use as heuristic

53. Learning Could also build pattern database while solving cases of the 8-puzzle Must keep track of intermediate states and true final cost of solutionMust keep track of intermediate states and true final cost of solution Inductive learning builds mapping of state -> costInductive learning builds mapping of state -> cost Because too many permutations of actual statesBecause too many permutations of actual states –Construct important features to reduce size of space Could also build pattern database while solving cases of the 8-puzzle Must keep track of intermediate states and true final cost of solutionMust keep track of intermediate states and true final cost of solution Inductive learning builds mapping of state -> costInductive learning builds mapping of state -> cost Because too many permutations of actual statesBecause too many permutations of actual states –Construct important features to reduce size of space